Systems and methods for parallel processing optimization for an evolutionary algorithm

Abstract

The systems and methods may include receiving an initial population of parent chromosome data structures, where each parent chromosome data structure provides a plurality of genes; selecting pairs of parent chromosome data structures; applying at least one evolutionary operator to the genes of the selected pairs to generate a plurality of child chromosome data structures; allocating, the generated plurality of child chromosome structures to a plurality slave processors, where each slave processor evaluates one or more of the plurality of child chromosome data structures and generates respective objective function values; receiving objective function values for a portion of the plurality of allocated child chromosome data structures; merging the parent chromosome data structures with the received portion of the child chromosome data structures for which objective function values have been received; and identifying a portion of the merged set of chromosome data structures as an elite set of chromosome data structures.

Description

RELATED APPLICATION[0001]The present application claims priority to U.S. Provisional Application Ser. No. 61 / 178,806, filed on May 15, 2009, and entitled “Systems and Methods for Utilizing Genetic Resources for Innovation and Problem Solving,” which is hereby incorporated by reference in its entirety as if fully set forth herein.FIELD OF THE INVENTION[0002]Aspects of the invention related generally to evolutionary algorithms and other genetic resources, and more particularly to systems and methods for utilizing evolutionary algorithms and other genetic resources to produce designs for multi-objective optimization problems.BACKGROUND OF THE INVENTION[0003]The goal of multiple-objective optimization, in stark contrast to the single-objective case where the global optimum is desired (except in certain multimodal cases), is to maximize or minimize multiple measures of performance simultaneously whereas maintaining a diverse set of Pareto-optimal solutions. The concept of Pareto optimali...

AI Technical Summary

Benefits of technology

[0004]According to an example embodiment of the invention, there is a method. The method may include receiving, by a master processor, an initial population of parent chromosome data structures, wherein each parent chromosome data structure provides a plurality of genes representative of variables in which associated values are permitted to evolve; selecting, by the master processor, pairs of parent chromosome data structures from the input population of parent chromosome data structures, applying, by the master processor, at least one evolutionary operator to the genes of the selected pairs of parent chromosome data structures to generate a plurality of child chromosome data structures; allocating, by the master processor, the generated plurality of child chromosome structures to a plurality slave processors, where each slave processor evaluates one or more of the plurality of child chromosome data structures with a plurality of objective functions and generates respective objective function values; receiving, by the master processor, objective function values for a portion of the plurality of allocated child chromosome data structures; merging, by the master processor, the parent chromosome data structures with the received portion of the child chromosome data structures for which objective function values have been received; and identifying, by the master processor, a portion of the merged set of chromosome data structures as an elite set of chromosome data structures based at least in part on the received objective function values, where the prior steps are performed for a plurality of iterations, including a first iteration and a second iteration, until termination criteria is satisfied, and where during a second iteration subsequent to the first iteration, the master processor receives objective function values for late child chromosome data structures associated with the one or more pairs of parent chromosome data structures allocated to slave processors during the first iteration, where during the second iteration, the elite set of chromosome data structures is utilized as the input population when performing the prior steps, and wherein during the second iteration, the late child chromosome data structures for which objective function values have been received are merged with the parent chromosome data structures that are associated with the input population of elite chromosome data structures.

[0005]According to

Problems solved by technology

Despite these advantages, real-world problems, such as satellite constellation design optimization and airline network scheduling optimization, challenge the effectiveness of classical methods.

These methods also limit discovery in the feasible solution space by requiring the decision maker apply some sort of higher-level information before the optimization is performed.

Images